What problem does it solve? End-to-end test suites become flaky when tests hardcode current content, share accounts across parallel workers, or create records nothing cleans up. This Skill establishes a single doctrine for how tests relate to data — discovery of data sources, a two-strategy decision ladder, and twelve rules covering generation, cleanup, and feasibility gating. ## Core Features & Use Cases - Two-strategy decision ladder: Classifies every data dependency as either seed-your-own-data (programmatic generation with hook-based cleanup) or content-resilient handling (runtime resolution of declared requirements against the app's own source). - Twelve operational rules: Covers per-attempt unique data generation, three-class dependency-error taxonomy (premise/app-state/infra), worker-scoped account pools, hydration-safe form filling, and a data-feasibility composing gate. - Living test data plan: Maintains tests/e2e/docs/test-data-plan.md recording each dependency's current strategy, gaps, and a roadmap to an ideal test environment. - Use Case: When composing a signup or order-creation test, apply the ladder to decide whether to seed data via API, resolve content at runtime, or block the scenario — then record the outcome in the test data plan. ## Quick Start Apply the test data conventions to decide how the new checkout test should source, isolate, and clean up its data.
